Speaking & Writing

You fire off a prompt to ChatGPT, then immediately wonder what Claude would say. So you open another tab, paste the same question, compare.

TL;DR: Want to extend your AI’s capabilities? The options are currently confusing.

I was reflecting on what I’ve written in the past year, as you sometimes do in this season. What a great time to be alive: to see the 70-year promise of AI start to be fulfilled.

TL;DR: Google just released Conductor, a Gemini CLI extension for structured AI development. It joins a growing family of tools (including Codev, Amazon’s Kiro, and GitHub’s Spec Kit) that all...

TL;DR: I’ve been talking to a lot of developers about how they use AI. What’s striking isn’t the tools they use, it’s how differently they work with them.

TL;DR: When bootstrapping a product, the scarcest resource isn’t money or talent. Every hour you spend on a problem that only matters if you succeed is an hour stolen from making success happen.

TL;DR: I was really into MCP, but that was before I discovered half my context window was gone before I’d typed a single instruction. It turns out MCP was solving the right problem (flexible,...

TL;DR: if you’re repeatedly pasting debug logs and asking the AI to fix it, there’s a better way: create a fixloop and remove yourself from the debugging cycle. A fixloop is where the AI can see if...

TL;DR: Currently, when using AI for software development, the typical work pattern is “Round Robin AI Pair Programming”: you have 1–3 sessions with AI running at a time and you alternate between...

TL;DR: The hype around imminent AGI is overblown. Current AI systems have fundamental weaknesses.

TL;DR: We’re still learning how to integrate chat interfaces in existing web apps. A lot of big companies are getting this badly wrong (including Google and Meta).

TL;DR: With AI agents writing our code, natural language has become the new high-level programming language. But we’re still treating documentation like an afterthought while throwing away the...

TL;DR: Advanced users should consider Claude Code as their daily driver AI, not just for coding. It’s the clearest glimpse we have of the agentic AI future: where agents actually have agency.

StockApp's engineering team designed their development culture around AI collaboration from inception. The team reports achieving approximately 2.

TL;DR**: If you’re feeling lost about your career trajectory, you’re in good company. From Prophet Muhammad to Mel Robbins, history shows us that meaningful careers often don’t reveal themselves...

I’ve wrestled for a long time about trying to say something insightful or novel about Israel, but two recent articles in The New York Times inspired me to take a different approach. Rather than add...

AI generated image of a robotic cat jumping through a catflap TL;DR: The “edit trick”, like many good ideas, is simultaneously obvious and not. When annotating documents using LLMs, don’t send the...

AI generated image of a lumberjack using a chainsaw blindfolded TL;DR: Vibe-coding is a horrible idea. It adds unnecessary complexity, it doesn’t always follow instructions, and sometimes it just...

TL;DR: MCP is really taking off. There are now thousands of MCP “servers”, and even though Anthropic invented it, just a few days ago OpenAI adopted it too.

Context ripples out in many directions. Generated using Flux Pro 1.

TL;DR: Claude Code is remarkable but requires intentional learning. I’ve used it for Python development, 3D design editing, documentation and even this article — outperforming alternatives across...

TL;DR: DeepSeek is a wonderful step in the development of open AI approaches. It also has a pretty serious pro-Chinese bias.

The release of OpenAI’s o1 model has confused many. Unlike previous models which were always about more (more parameters!

ICML is one of the premier conferences in AI, perhaps only behind NeurIPS in prestige. It’s a great place to understand what’s happening at the cutting edge of AI.

TL;DR: To get the most out of interviews, use the answers to your questions so far to guide your next question. Make your questions up on the fly.

Generated by Dall-E 3 for the prompt “Generate an image appropriate for an Islamic AI assistant called Ansari” Over the last year, I’ve been working on an Islamic AI assistant called Ansari. The...

I recently started experimenting with how LLMs (in particular GPT-3. 5-turbo) answer questions about Islam.

\[NOTE: This is my personal opinion and does not represent the views of my employer\] “It’s one of these unfortunate consequences. No one at Facebook is malevolent, but the incentives are misaligned.

This post examines how production ML architectures have evolved through distinct generations, drawing parallels to GPU programming history. First Generation: Featured "fixed function" pipelines with...

Ray is a general purpose distributed computing paradigm that has been gaining momentum in both open source and commercial circles. People are using it for a wide variety of tasks: from a substrate...

This blog post addresses the lack of reproducible benchmarking in LLM inference claims. The team open-sourced LLMPerf, a tool designed to standardize performance measurement across different LLM...

Today we're announcing Aviary, a new open source project that simplifies and enables easy self-hosted serving of multiple LLM models efficiently. Open source LLMs are getting better every day, so...

This guide provides essential metrics for LLM developers to use in technical planning and cost estimation. The piece draws inspiration from Jeff Dean's famous "Numbers every Engineer should know"...

This blog post demonstrates how to construct a semantic search engine using LangChain and Ray in approximately 100 lines of code. I walk through building an index of Ray's documentation and making...

I'm excited to announce that I'm joining Anyscale, a company focused on simplifying distributed computing. Reflecting on my 20-year career working with distributed systems at Google and Uber, I...

Hi, my name's Ollie Kadus and today I'd like to demo agent farm to you. Agent farm is the UI for the codev project, a project that's an operating system for how humans and AI should build software...

Channel: Cluesmith LLC

A comprehensive 6-hour hands-on workshop empowering Imams and Muslim community workers to integrate AI tools into their daily work: khutbah preparation, community outreach, research, translation, and...

Channel: AMJA

You're listening to Wise Women with Daisy Khan, where faith rolls up its sleeves as a force for good and where women define the future with wisdom and wonder. I'm Daisy Khan, author, speaker, and a...

Channel: Wise Women with Daisy Khan

Muhammed [Music] Doctorchee. Uber, Google I'm sorry.

Channel: Quran Computing

[Music] our next speaker Dr red do will share his experiences building and running and S chat and will discuss the opportunities and risks of associated with integrating Ai and Islam Dr will lead is...

Channel: IQRA

I still think that that's a very difficult problem real-time sensing real-time interaction you know any any human you know with 10 minutes of training can probably fold laundry yet there isn't a...

Channel: Startups Arabia Podcast

Uh thank you so much uh for the kind words at the beginning. Um uh I want to welcome everybody who came here especially the people who came from interstate and one of my favorite hadith is the...

Channel: Masjid DarusSalam

[Music] hello and welcome to the iot chat where we explore the latest developments in the internet things I'm your host Christina Cardoza editorial director of inside. te and today we're going to be...

Channel: insightdottech

our next speaker is the chief scientist at any scale and as uh as uh Sadir mentioned we are on in mile one many of us are running experiments figuring out how to use uh llms and Foundation models...

Channel: Industrial AI Federation

all right so today we're in for a treat we have wed kadus who is the chief scientist at any scale and again full disclosure I am an adviser to any scale and for our listeners this is an episode that...

Channel: Gradient Flow

thank you very much for joining me today it's a pleasure to speak to you all um so I I'm going to talk a little bit about uh what it takes to get llms from demo to production and so why don't I start...

Channel: The AI Conference

[Music] thank you hello my name is Waleed cadus I'm the chief scientist here at NHL and I'm really excited to talk to you today about going Beyond prototypes it's been very easy with chat GPT and...

Channel: AI Infrastructure Alliance

[Music] foreign [Music] thank you so much for joining us um good morning I think it's morning from where you're joining us from right something like that midday yes good day all right then so I guess...

Channel: 3logy Limited

Channel: OnePath Network

I'm not in ASMR mode anymore you there what's up Waleed hey how's it going dimitrios are you doing all right I'm doing great man I'm blissing out now after that one oh yeah there we go there we go...

Channel: MLOps.community

yeah my name is Willie cadus I'm the head of engineering here at any scale a company that is the behind the open source project right and believe it or not I really don't drink coffee maybe I'm the...

Channel: MLOps.community

[Music] um it's an absolute pleasure to be with you folks again i had such a blast last time with dee um at the last year's applied conference and uh i'm really grateful for the introduction i i had...

Channel: Tecton

[Music] so [Music] hello and welcome to today's wief think talk series 2021 final edition titled ai and machines learning for a competitive edge now let's welcome our solo performance by a wief...

Channel: WIEF FoundationTV

[Music] excellent so it's really great to see you folks today um i'm really excited to share some thinking that i've been doing recently about where is where are production ml architectures going let...

Channel: Tecton

My name is Wed Kadus and with me is Mark Stitus. We're members of the Android location and context team and we're super excited to be talking to you today about how to make your apps context aware.

Channel: Google for Developers

Channel: Google for Developers

okay thank you very much we're really excited that you've joined us today my name is w cadus and next to me is Sarah Peterson the business product manager for indoor maps and we're really excited to...

Channel: Google for Developers

This presentation discusses the development of Ansari Chat, an AI assistant designed to help Muslims access Islamic knowledge and scholarship. Beginning as a basic chatbot in 2023, Ansari has grown...

In the Name of Allah, Most Beneficent, the Most Merciful By the end of this session: Realize criticality of Ummah adopting AI Get you actually using AI to be more creative, productive, knowledgeable,...

The Prophetic Strategy Summit - Putrajaya, Malaysia

Ansari (https://ansari. chat) is a digital assistant for Muslims built on LLMs.

What are the "use case patterns" for deploying LLMs into production? Understanding these will allow you to spot "LLM-shaped" problems in your own industry.

Are Open LLMs useful for production applications, or are they low quality toys useful only for experiments? We share our experiences using open LLMs vs proprietary LLMs.

Reinforcement Learning has historically not been as widely adopted in production as other learning approaches (particularly supervised learning), despite being capable of addressing a broader set of...

Production ML architectures (deployed at scale in production) are evolving at a rapid pace. We suggest there have been two generations so far: the first generation were very much fixed function...

Published as a conference paper at ICLR 2025 ROUTE LLM: L EARNING TO ROUTE LLM S WITH P REFERENCE DATA Isaac Ong∗1 Amjad Almahairi∗2 Vincent Wu1 Wei-Lin Chiang1 Joseph E. Gonzalez1 M Waleed Kadous3...

This article features an interview with Dr. Waleed Kadous, a computer scientist who has worked on AI integration for major tech companies like Google and Uber.